[time-nuts] Need advice for multilateration setup

Magnus Danielson magnus at rubidium.dyndns.org
Mon Apr 6 12:57:04 EDT 2015


On 04/06/2015 03:13 PM, Jim Lux wrote:
> On 4/6/15 2:21 AM, Attila Kinali wrote:
>> On Sat, 04 Apr 2015 08:49:01 +0200
>> Magnus Danielson <magnus at rubidium.se> wrote:
>>> This is on either side of the amateur 23 cm band. That's also the first
>>> band where you have bandwidth enough to fool around with stuff like this
>>> without breaking the bandplan.
>> This shouldn't be much of a problem. Using a chiping rate of a couple
>> of kHz should be enough for this application. The signal strength
>> can be rather large, directive antennas can be used and the expected
>> noise level is rather low. So there no need to use a high chipping
>> rate to compensate for noise effects. Of course, using a higher
>> chipping rate makes it also easier to get an higher accuracy, but
>> I would start with something easy to do first, like a 100mW transmitter
>> in the 70cm band with 10kHz chipping rate (or go to a sub-band,
>> where 200kHz signals are allowed). With that kind of setup it should
>> be possible to use something like RTL-SDR for the first experiments
>> and then gradually upgrade to better hardware to improve accuracy.
> One strategy for this kind of application is to do the "fine
> measurement" using carrier phase, and use the PN code to do ambiguity
> reduction.  Then, a low chip rate is fine: you're basically using it as
> a check that you haven't "slipped a cycle".
> I would think that the RTL dongles would work just fine, especially if
> you radiate a pilot tone from a fixed location as well as the tone from
> the rocket.  You basically set up two PLLs in software one to track each
> tone, and subtract the phase of one from the other for each ground station.

Indeed. Considering that in the start location you can solve integer 
ambiguity, especially with a pilot-tone or several radiated. Hacking in 
on the RTL-SDR to steer or replace the clock with a more stable clock 
might be considered.


More information about the time-nuts mailing list